Oxygen Overlayer Research Articles

First-principles studies on clean and oxygen-adsorbed Ir(110) surfaces

Using density functional theory in combination with an ab initio atomistic thermodynamics approach the structure and stability of Ir(110) surfaces in contact with an oxygen atmosphere have been studied. Besides the unreconstructed surface, $(1\ifmmode\times\else\texttimes\fi{}2)$-, $(1\ifmmode\times\else\texttimes\fi{}3)$-, and $(1\ifmmode\times\else\texttimes\fi{}4)$-reconstructed Ir(110) have been considered. We find that without adsorption of oxygen all reconstructed surfaces are more stable than $\mathrm{Ir}(110)\text{\ensuremath{-}}(1\ifmmode\times\else\texttimes\fi{}1)$. Adsorption of oxygen with coverages higher than 0.5 ML removes the surface reconstruction, leading to either a $c(2\ifmmode\times\else\texttimes\fi{}2)$ or $p(2\ifmmode\times\else\texttimes\fi{}1)$ oxygen overlayer on the unreconstructed Ir(110) surface and therefore an adsorbate-induced lifting of the reconstruction. While on almost all surfaces oxygen prefers binding at bridge positions to Ir atoms of the topmost layer, on the reconstructed surfaces higher coverages are required to additionally stabilize oxygen at adjacent threefold sites. Besides the stability of surface structures, our studies also reveal quantitative information on the geometry and energetics, providing further insights into the sometimes controversial discussion of oxygen adsorption on Ir(110).

Structure of the Hydrated (101̄4) Surface of Rhodochrosite (MnCO3)

The three-dimensional structure of the hydrated (1014) surface of MnCO3 at 90% relative humidity and 295 K is determined from measurements of X-ray scattering along ten crystal-truncation rods (CTRs). The scattering data provide both vertical and lateral information about the interfacial structure. The model that best fits the scattering data is a surface having a first layer of manganese carbonate and an overlayer of oxygen (as water). Within the measurement uncertainty, the overlayer of oxygen (O(w)) and the first-layer of manganese (Mn1) have equal occupancies of 0.84. The Mn1-O(w) distance between these layers is 2.59 +/- 0.04 angstroms. The overlayer O atoms are displaced laterally by 0.157 angstroms in the x- and 0.626 angstroms in the y-direction relative to the first-layer Mn atoms. The first-layer carbonate groups tilt by -4.2 +/- 2.1 degrees in phi (toward the surface plane) and -2.6 +/- 1.2 degrees in chi (an axis perpendicular to phi). The second-layer carbonate groups do not tilt, at least within measurement uncertainty. The spacing between Mn atom layers remains unchanged within measurement error whereas the spacing between layers of C atoms in carbonate contracts for the top three layers. Knowledge of the detailed atomic structure of the hydrated (1014) surface of MnCO3 provides a structural baseline for the interpretation of chemical reactivity.

Formation and hydrogenation of p(2 × 2)-N on Pt(1 1 1)

The formation of a well-ordered p(2 × 2) overlayer of atomic nitrogen on the Pt(1 1 1) surface and its reaction with hydrogen were characterized with reflection absorption infrared spectroscopy (RAIRS), temperature programmed desorption (TPD), low energy electron diffraction (LEED), Auger electron spectroscopy (AES), and X-ray photoelectron spectroscopy (XPS). The p(2 × 2)-N overlayer is formed by exposure of ammonia to a surface at 85 K that is covered with 0.44 monolayer (ML) of molecular oxygen and then heating to 400 K. The reaction between ammonia and oxygen produces water, which desorbs below 400 K. The only desorption product observed above 400 K is molecular nitrogen, which has a peak desorption temperature of 453 K. The absence of oxygen after the 400 K anneal is confirmed with AES. Although atomic nitrogen can also be produced on the surface through the reaction of ammonia with an atomic, rather than molecular, oxygen overlayer at a saturation coverage of 0.25 ML, the yield of surface nitrogen is significantly less, as indicated by the N 2 TPD peak area. Atomic nitrogen readily reacts with hydrogen to produce the NH species, which is characterized with RAIRS by an intense and narrow (FWHM ∼ 4 cm −1) peak at 3322 cm −1. The areas of the H 2 TPD peak associated with NH dissociation and the XPS N 1s peak associated with the NH species indicate that not all of the surface N atoms can be converted to NH by the methods used here.

Efficient CO Oxidation at Low Temperature on Au(111)

The rate of CO oxidation to CO2 depends strongly on the reaction temperature and characteristics of the oxygen overlayer on Au(111). The factors that contribute to the temperature dependence in the oxidation rate are (1) the residence time of CO on the surface, (2) the island size containing Au-O complexes, and (3) the local properties, including the degree of order of the oxygen layer. Three different types of oxygen--defined as chemisorbed oxygen, a surface oxide, and a bulk oxide--are identified and shown to have different reactivity. The relative populations of the various oxygen species depend on the preparation temperature and the oxygen coverage. The highest rate of CO oxidation was observed for an initial oxygen coverage of 0.5 monolayers that was deposited at 200 K where the density of chemisorbed oxygen is maximized. The rate decreases when two-dimensional islands of the surface oxide are populated and further decreases when three-dimensional bulk gold oxide forms. Our results are significant for designing catalytic processes that use Au for CO oxidation, because they suggest that the most efficient oxidation of CO occurs at low temperature--even below room temperature--as long as oxygen could be adsorbed on the surface.

Mechanism of CO oxidation reaction on O-covered Pd(111) surfaces studied with fast x-ray photoelectron spectroscopy: Change of reaction path accompanying phase transition of O domains

We studied the mechanism of CO oxidation on O-precovered Pd(111) surfaces by means of fast x-ray photoelectron spectroscopy (XPS). The oxygen overlayer is compressed upon CO coadsorption from a p(2 x 2) structure into a (square root(3) x square root(3))R30 degrees structure and then into a p(2 x 1) structure with increasing CO coverage. These three O phases exhibit distinctly different reactivities. (1) The p(2 x 2) phase does not react with CO unless the surface temperature is sufficiently high (<290 K). (2) In the square root(3) x square root(3))R30 degrees phase, the reaction occurs exclusively at island peripheries. CO molecules in a high-density phase formed under CO exposure react with oxygen atoms, leading to quite a small apparent activation energy. (3) The reaction proceeds uniformly over the islands in the p(2 x 1) phase.

Interaction of Ag and O on W(1 1 0) studied by scanning tunneling microscopy

The structural changes of Ag films on W(1 1 0) upon coadsorption of oxygen have been studied by scanning tunneling microscopy. The exposure of one monolayer Ag to oxygen leads to a phase separation into an Ag bilayer and patches of O-covered W(1 1 0). The effective Ag island thickness increases linearly with oxygen exposure. For Ag submonolayer-islands the onset of the bilayer formation is delayed, the induction period increases with the available free W area. We conclude that the steps of the transport process are (1) dissociation of oxygen on W and on the Ag islands, (2) site exchange of atomic oxygen with Ag atoms predominantly at the island edges – while on W(1 1 0) the oxygen is immobile, (3) diffusion of the displaced Ag atoms to the island edges where they are incorporated into the monolayer and (4) initiation of Ag bilayer formation, once the W(1 1 0) is saturated with O. This indicates an unexpected activity of the Ag monolayer on W(1 1 0) towards oxygen dissociation. In case of a reversed deposition sequence, where submonolayer quantities of Ag are adsorbed on an oxygen-precovered W(1 1 0) surface, growth of Ag clusters is observed. The distribution of cluster size and cluster height depends critically on the spatial order within the predeposited oxygen overlayer – it is obvious that the oxygen overlayer on the W surface acts as a structured template for preferential Ag nucleation.

Theoretical study of oxygen adsorption at the Fe(1 1 0) and (1 0 0) surfaces

Electronic, magnetic and structural properties of atomic oxygen adsorbed in on-surface and subsurface sites at the two most densely packed iron surfaces are investigated using density functional theory combined with a thermodynamics formalism. Oxygen coverages varying from a quarter to two monolayers (MLs) are considered. At a 1/4 ML coverage, the most stable on-surface adsorption sites are the twofold long bridge sites on the (1 1 0), and the fourfold-hollow sites on the (1 0 0) surface. The presence of on-surface oxygen atoms enhances the magnetic moments of the atoms of the two topmost Fe layers. Detailed results on the surface magnetic properties, due to O incorporation, are presented as well. Subsurface adsorption is found unfavored. The most stable subsurface O, in tetrahedral positions at the (1 0 0) and octahedral ones at the (1 1 0) surface, are characterized by substantially lower binding than that in the on-surface sites. Subsurface oxygen increases the interplanar distance between the uppermost Fe layers. The preadsorbed oxygen overlayer enhances binding of subsurface O atoms, particularly for tetrahedral sites beneath the (1 1 0) surface.

The adsorption of and reaction of NO 2 on Ag(1 1 1)-p(4 × 4)-O and formation of surface nitrate

The adsorption and reaction of nitrogen dioxide on the Ag(1 1 1)-p(4 × 4)-O surface has been investigated with RAIRS, TPRS and STM. At 300 K NO 2 initially reacts with the oxygen overlayer to form nitrate in p(3 × 3) and p(4 × 4) structures, which convert to a new p(3 × 3) at saturation coverage. Surface pitting during nitrate adsorption is suggestive of the incorporation of silver atoms into the NO 3 structure. With heating NO 3 decomposes into NO 2 and O at 396 K and 497 K, and oxygen desorbs at 578 K.

Density-functional study of oxidation at the Mn–Co interface

By using first-principles plane-wave pseudopotential code, with the Perdew–Wang 91 form of the density functional, we study the Mn–Co magnetic coupling of ultrathin Mn films on fcc Co(001) substrates and the effect of oxygen on this Mn–Co coupling. First, in agreement with experiments and recent calculations we find the MnCo (2×2) alloy overlayer with ferromagnetic coupling of all spins as the most stable structure among several other natural candidates. The magnetic coupling remains however unchanged after deposition of a full oxygen overlayer, leaving thus unexplained the experimentally observed Mn antiferromagnetic moment alignment. For more complicated structures, however, indications of antiferromagnetic stabilization appear. Particularly, the antiferromagnetic ground state of a model two-layer (Mn,Co)O oxide slab with rock-salt structure on Co(001) is found more stable by 0.14eV per elementary cell than the ferromagnetic state. These results display clear indication that superexchange in bimetallic Mn–O–Co groups can stabilize the Mn–Co antiferromagnetic coupling.

Structural and kinetic effects on a simple catalytic reaction: Oxygen reduction on Ni(110)

Oxygen hydrogenation at 100 K by gas phase atomic hydrogen on Ni(110) has been studied under ultrahigh vacuum conditions by temperature programmed desorption (TPD) and x-ray photoelectron spectroscopy (XPS). Formation of adsorbed water and hydroxyl species was observed and characterized. The coverage of the reaction products was monitored as a function of both temperature and initial oxygen precoverage. On the contrary, when high coverage oxygen overlayers were exposed to gas phase molecular hydrogen, no hydrogenation reaction took place. The results are compared to the inverse process, exposing the hydrogen covered surface to molecular oxygen. In this case, at 100 K, simple Langmuir-Hinshelwood modeling yields an initial sticking coefficient for oxygen adsorption equal to 0.26, considerably lower than for the clean surface. Moreover, formation of hydroxyl groups is found to be twice as fast as the final hydrogenation of OH groups to water. Assuming a preexponential factor of 10(13) s(-1), an activation barrier of 6.7 kcal/mol is obtained for OH formation, thus confirming the high hydrogenating activity of nickel with respect to other transition metals, for which higher activation energies are reported. However, oxygen is hardly removed by hydrogen on nickel: this is explained on the basis of the strong Ni-O chemical bond. The hydrogen residual coverage is well described including a contribution from the adsorption-induced H desorption process which takes place during the oxygen uptake and which is clearly visible from the TPD data.

Oxygen Overlayers on Pd(111) Studied by Density Functional Theory

By use of density-functional theory we analyze the on-surface adsorption of oxygen on Pd(111) for coverages up to 1 monolayer and compare the results with corresponding data for the other late 4d transition metals, namely, Ru, Rh, and Ag. Besides the known effect of the continued d-band filling on the oxygen-metal bond strength, we also discern trends in the adsorption geometries, work functions, and electron density of states. The repulsive lateral interactions in the overlayer give rise to a pronounced reduction of the adsorption

Substrate Dependence of Self-Assembly of Alkanethiol: X-ray Absorption Fine Structure Study

We have investigated effects of substrates on self-assembly of alkanethiol molecules by means of X-ray absorption fine structure spectroscopy. Molecular orientation of hexanethiolate adsorbed on various substrates such as Cu(111), Ag(111), Ag(100), Si(111), and their oxygen-covered surfaces has been explored by C K-edge and S K-edge near-edge X-ray absorption fine structure (NEXAFS). It was revealed that the alkyl-chain orientation is not correlated to the S−C bond orientation but primarily determined by the molecular density that definitely depends on the substrate surface. Precovered oxygen overlayers do not prevent thiol molecules from being adsorbed but promote their adsorption rate. The surface oxygen atoms are removed via formation and desorption of H2O. The interface structure of a hexanethiolate monolayer on O/Cu(100) has been studied in detail by use of S K-edge surface extended X-ray absorption fine structure (SEXAFS). The SEXAFS results indicated that hexanethiol adsorption induces a lateral displacement of surface Cu atoms toward the sulfur atom of the thiolate with a concomitant breaking of the −O−Cu−O− one-dimensional chain.

The reaction of CH3NO2 on O-covered Mo(110): the effect of oxygen on product distribution

The reactions of CH 3 NO 2 on Mo(1 1 0) containing 0.66 and 0.40 ML of chemisorbed oxygen are studied using temperature programmed reaction (TPR) and reflection absorption infrared spectroscopy (RAIRS). Nitromethane decomposes on the surface with 0.66 ML of O––occupying both twofold and quasi-threefold coordinated sites––below 300 K via dissociation of N–O bonds to form methylimido, CH 3 N(a). The oxygen is deposited on the surface populating terminal sites while the methylimido yields gas-phase HCN, H 2 O, and methyl radical between ∼550 and 800 K. In addition, some methylimido decomposes nonselectively to adsorbed N and C. The reaction on the 0.40 ML surface (containing only twofold coordinated oxygen) also affords HCN, H 2 O, CH 3 , and adsorbed O, N, and C. The product distribution is different when the initial oxygen coverage is 0.4 ML––the amount of HCN and H 2 O is ∼10% greater, the yield of CH 3 is ∼10% less, and the amount of adsorbed C and O is substantially increased. In addition, the desorption temperature of the C–H bond scission products is ∼20 K lower on the 0.4 ML oxygen surface than on the saturated surface. The reactions on the chemisorbed oxygen overlayers are compared to those of CH 3 NO 2 on a thin film oxide of Mo(1 1 0). The observed differences in reactivity as a function of oxygen coverage and as a function of the presence of subsurface oxygen are discussed in terms of surface preparations which favor either C–H bond scission or oxygenate evolution.

The reaction of CH 3NO 2 on O-covered Mo(1 1 0): the effect of oxygen on product distribution

The reactions of CH 3NO 2 on Mo(1 1 0) containing 0.66 and 0.40 ML of chemisorbed oxygen are studied using temperature programmed reaction (TPR) and reflection absorption infrared spectroscopy (RAIRS). Nitromethane decomposes on the surface with 0.66 ML of O––occupying both twofold and quasi-threefold coordinated sites––below 300 K via dissociation of N–O bonds to form methylimido, CH 3N(a). The oxygen is deposited on the surface populating terminal sites while the methylimido yields gas-phase HCN, H 2O, and methyl radical between ∼550 and 800 K. In addition, some methylimido decomposes nonselectively to adsorbed N and C. The reaction on the 0.40 ML surface (containing only twofold coordinated oxygen) also affords HCN, H 2O, CH 3, and adsorbed O, N, and C. The product distribution is different when the initial oxygen coverage is 0.4 ML––the amount of HCN and H 2O is ∼10% greater, the yield of CH 3 is ∼10% less, and the amount of adsorbed C and O is substantially increased. In addition, the desorption temperature of the C–H bond scission products is ∼20 K lower on the 0.4 ML oxygen surface than on the saturated surface. The reactions on the chemisorbed oxygen overlayers are compared to those of CH 3NO 2 on a thin film oxide of Mo(1 1 0). The observed differences in reactivity as a function of oxygen coverage and as a function of the presence of subsurface oxygen are discussed in terms of surface preparations which favor either C–H bond scission or oxygenate evolution.

Modification of dehydrogenation activity of nickel on O-covered Mo(1 1 0)

The reactivity of methanol on Ni ( θ⩽1.2 ML) on Mo(1 1 0)–(1 × 6)-O is studied using temperature programmed reaction and X-ray photoelectron spectroscopies. Carbon monoxide, D 2O, D 2, and formaldehyde are evolved from CD 3OD. The formaldehyde production represents a departure from reactivity typical for nickel, which usually promotes complete dehydrogenation. Similarly, formaldehyde production is atypical of Mo(1 1 0)–(1 × 6)-O. Isotopic labeling studies using CH 3OH, CD 3OD, CD 2HOH and Mo(1 1 0)–(1 × 6)- 18O show that formaldehyde is produced from irreversible C–H(D) bond dissociation of methoxy. Furthermore, there is a significant kinetic isotope effect in production of formaldehyde, such that yields from the CH 3OH reaction are ∼4 times less than formaldehyde yields from the CD 3OD reaction. Differences in the satellite structure of the Ni(2p) peaks for Ni deposited on Mo(1 1 0)–(1 × 6)-O compared to Ni on clean Mo(1 1 0) provide evidence for subtle perturbations of the electronic structure of the Ni deposited on the oxygen overlayer, possibly associated with formation of a strained Ni layer. X-ray photoelectron data also rule out the possibility that the difference in reactivity for Ni deposited on clean vs. the (1 × 6)-O overlayer is due to oxygen migration from the Mo to the Ni. Finally, the variation in formaldehyde yields as a function of Ni coverage is used to rule out special edge or defect sites as the reaction centers leading to formaldehyde production. These results suggest that interfacial interactions can be used to modify the electronic structure of metals so as to alter reaction selectivity.

Mesoscopic models of oxygen migration on the Ru(001) surface

Atomic migration within the organized oxygen overlayers on the basal plane of Ru(001) is modeled within phenomenological mesoscopic model accounting for all possible atomic jumps within the adsorbate and capable of accessing several time scales in the migration kinetics. Experimentally observed diffusion and trapping of the isolated oxygen vacancy within the 30(2x2) overlayer at coverage θ=3/4 are accounted for. The comparison with experiment is used to propose values of parameters for interactions between an atom adsorbed at a hcp site and another one temporarily at some of the neighboring bridge sites. The modeling is extended to the O(2×2) overlayer at θ=1/4 and it is shown that for this structure a competing diffusion channel exists which breaks down the equivalence of the atomic migration behavior between the organized θ=1/4 and 0=3/4 overlayers. For both overlayers simplified effective models are formulated, free of unobservable metastable configurations and unobservable time scales. Considerations for the ideal O(2×1) rowlike overlayer at θ=½ confirm that occasional jumps of the oxygen atoms to one side of the rows should be observable in the STM images for this overlayer in contrast to the θ= 1/4 and 3/4 overlayers for which an excess atom or a vacancy, respectively, are necessary to initiate observable atomic migration processes.

The structure of disordered chemisorbed oxygen on Cu [formula omitted

The structure of a disordered chemisorbed oxygen layer on Cu(1 1 1) has been investigated using normal incidence X-ray standing wavefield absorption (NIXSW). The NIXSW data indicates that the oxygen atoms within the overlayer have more than one well-defined height with respect to the copper surface. The NIXSW data has been compared to previous proposed structural models for the disordered overlayer. We believe that the model that is most consistent with our data is the oxygen overlayer having a Cu 2O-like structure.

Role of long range interaction in oxygen superstructure formation on Cu(001) and Ni(001).

On the basis of electronic structure calculations, we show that the long range Coulomb interaction provides the driving mechanism for oxygen overlayer formation on Cu(001). We illustrate that this interaction in the precursor c(2 x 2) phase induces a missing row reconstruction of Cu(001), and leads to the (2sqrt[2] x sqrt[2])R45 degrees O structure, which has strong covalent pO-dCu coupling. For the c(2 x 2)O overlayer on Ni(001) and Cu(001), we show that pO-dNi bonding is larger than pO-dCu and serves to neutralize the perturbation of the Coulomb interaction induced by the O overlayer. Consequently, c(2 x 2)O/Ni(001)) is stable while c(2 x 2)O/Cu(001) exists only in limited environments.

Reaction of sulfur dioxide with Ag( [formula omitted])–p(2×1)-O: a LEED, TPRS, and STM investigation

STM has been combined with LEED, TPRS, XPS, and EELS to further the understanding of the chemical and structural changes accompanying the reaction of SO 2(g) with oxygen adsorbed on Ag(1 1 0). At 300 K sulfur dioxide reacts with the p(2×1) oxygen overlayer to yield a surface covered by sulfite in a c(6×2) arrangement with six sulfite species per unit cell; i.e., SO 2(g)+O(a)→SO 3(a). This reaction appears to involve rearrangement of the “added” silver in the original p(2×1)-O structure to precipitate new islands of sulfite-covered surface on top of the original surface. Overall, sulfite disproportionates according to 6SO 3(a)→2SO 2(g)+4SO 3(a)+2O(subsurface) upon heating to 500 K. This heating yields a surface with a p(1×1) LEED pattern, which is shown to consist of irregular, narrow rows of c(3×2) unit cells aligned 22° from the [0 0 1] direction, separated by domain boundaries. Continued heating to 600 K results in the further loss of SO 3(a), yielding a surface covered with sulfate in a p(3×2) structure, with one sulfate moiety per unit cell. The results of these experiments suggest the following surface reaction for the conversion of sulfite to sulfate: 6SO 3(a)→4SO 2(g)+2SO 4(a)+2O(subsurface). A detailed two-dimensional depiction of the surface with substrate registry is proposed.

Step-modified phase diagram of chemisorbed oxygen on nickel

We have studied the effect of an extended array of steps on the two-dimensional phase behavior of chemisorbed oxygen overlayers on a vicinal nickel surface using low energy electron diffraction (LEED), Auger electron spectroscopy, and scanning tunneling microscopy. Phase behavior of oxygen on the vicinal Ni(9 7 7) surface was examined and compared with that for oxygen adsorbed on the flat Ni(1 1 1) surface. There are two significant differences in the phase diagrams for these two surfaces. On Ni(1 1 1) at θ=0.25 ML, oxygen forms a p(2×2) structure that disorders to a lattice gas at 440 K and remains disordered until it is ultimately dissolved into the bulk above 500 K. Surface defects, such as the steps on Ni(9 7 7), substantially modify this phase progression. On Ni(9 7 7), the p(2×2) phase still disorders at 440 K, but a second ordered phase, which can be designated as Ni[ 8(1 1 1)×(1 0 0) ]–2(1d)-O in microfacet notation, exists between room temperature and above 500 K when the oxygen is finally incorporated into the bulk. This adsorbate phase is step-stabilized and can be generated by dosing the surface with a small amount of oxygen or as a result of partial dissolution of oxygen from the higher coverage p(2×2) phase. Moreover, anisotropic disordering effects are evident due to the presence of the steps as indicated by the increasingly oblate shape of diffraction spots as the p(2×2) disorders. The process of oxygen dissolution is also qualitatively altered by the presence of regular steps.